Select languages

  • Polish
  • English
 
 
Print E-mail

UNCERTAINTY OF RADIOCARBON DATE


Uncertainty of radiocarbon date quoted by 14C laboratory usually comprises uncertainties connected with 14C counting statistics, and with fluctuating level of contamination introduced during sample preparation, and fluctuating parameters of 14C- measurement.

 

Typical precision of our 14C dates

Uncertainty (error, precision) of 14C date (1-sigma) quoted by the Poznań Radiocarbon Laboratory is determinted basing on counting statistics or standard deviation of results (5 at least) of partial 14C measurements of  the dated sample, and of the samples of modern standard and background. Practically, significant deviations between these two estimates occur very rarely.

Average precision of 14C date depends on the age of sample. For young samples, the precision is mainly limited by 14C counting stastics, and when very many 14Cs are counted - by reproducibility of sample preparation. For very old samples, the crucial component of date uncertainty is fluctuation of background level, depending on the AMS spectrometer itself, and on contamination introduced during sample preparation. 
Uncertainty of individual 14C date may differ from the average. The statistics of uncertainties of 14C dates obtained for samples containing >1 miligram of carbon, is presented in diagrams shown below. Uncertainties of dates of samples smaller than 1 mgC increase with decreasing mass of sample

error1.png
error2.png
Statistics of uncertainty of 14C dates quoted by the Poznań Radiocarbon Laboratory. Distributions of uncertainty are different for different intervals of sample age. The presented statistics concerns samples containing >1mg of carbon


Small samples

Routinely, for 14C AMS measurement we forward 1 miligram of carbon. To analyse smaller samples, we studied the dependence of background on sample mass (using many small background samples), and studied efectiveness of correction for isotopic fractionation. These studies showed increasing scatter of 14C/12C ratios with decreasing mass of background samples. They also showed decrease of 13C/12C ratio with decreasing sample mass and demonstrated, that fractionation correction normally used in 14C measurements was not always effective for samples smaller than ca. 0.2 mg C. 

The error of dates increases with decreasing mass. For young samples, it is mostly due to uncertainty of isotopic fractionation correction. For old samples, uncertainty of background and low number of 14C counts becomes most critical. Average errors of our 14C dates, for different sample ages and sample masses, is shown in table displayed below.

For very small samples, we from time to time fail to make 14C measurement. Frequency of so unlucky measurements (denoted in the table as "failed") increases with decreasing mass of carbon (in the rage 0.1-0.2 mgC we fail to date ca. 10%  of samples)

Average uncertainties of 14C dates reported by the Poznań Radiocarbon Laboratory, shown for different sample ages and different masses of carbon
  Age (kyr BP)
 
failed
  0-1 1-3
3-6 6-10
10-15
15-20
20-25 25-30
30-35
35-40
40-45
45-50
  Mass (mg C)
 
 %
 
± average error (yrs )
  >1.0
 
---
  30
35
40
50
60
100
270
340
520 
660 
1200 
2000 
   0.8-1.0  
---
  30 35 40
50
60
110
270 360
550
700
1300
3000
   0.6-0.8  
---
  35 35
40
60
70
140
290 400
600
800
1500
----
   0.4-0.6  
---
  40 40
50
70
100
140
290
400
600
800
2000
4000
   0.3-0.4  
0.5%
  40
50
50
70
100 160
300
400
900
1000 3000 5000
   0.2-0.3
 
  2%
  50
50
60
80
110
170
300
410
1000 1300 3000
----
   0.1-0.2  
 10%
  60
60
70
100 150
200 370
530
1200
2000
4000
----
   0.0-0.1  
  51%
  100 110 110
160
190
250
360
900
----
----
----
----
 
© 2017 Poznańskie Laboratorium Radiowęglowe
Joomla! is Free Software released under the GNU/GPL License.